To create effective distributed service systems, you must be able to test the interaction of components at each stage of development. Often, such testing is carried out after the publication of the application --- on real users. In this paper, we will describe how to emulate and test a distributed system on the limited resources of each specific developer. What to do with factors that are beyond the scope of tests of one application? How will the system behave when: a network with high latencies is used, the service in the chain does not work, the subnet is unavailable? We have created a tool that allows you to monitor the logic of the service system when various events occur on the network at the level of abstraction, accessible to the developer without a deep understanding of the work of network infrastructures. The MADT product we created can be used to compare the operation of several distributed applications under the same conditions, to verify the stability of an application against unstable network ope